1	The Path to Oracle Fusion Dr. Paul Dorsey Dulcian, Inc. www.dulcian.com
2	Who am I? Author of lots of books JDeveloper, Database Design, PL/SQL, Designer, Developer Never worked for Oracle Oracle Fusion Middleware Director President Emeritus NYOUG President Oracle consulting/training company Business rules approach Business Rules Information Manager (BRIM^®) product architect “Dorsey’s a jerk.” “Dorsey’s still a jerk.”
3	What you will get out of this presentation (This is NOT an Oracle approved message) All of the different things that “Fusion” means Oracle is TOTALY committed to making parts of Fusion work. What parts of Fusion are worth learning about DBMS, OAS, ADF BC, ADF Faces Which parts can be ignored BPEL, BAM, Oracle Business Rules Don’t count on the transition being easy.
4	Fusion: What is it? In addition to e-Business, Oracle has purchased: PeopleSoft (previously bought JD Edwards) Siebel Retek AND… TimesTen, Context Media, G-Log, Oblix, TripleHop, ProfitLogic, i-flex, Innobase, Thor Technologies, TempoSoft, OctetString, 360Commerce, Sleepycat, HotSip, Portal Software, Demantra, Net4Call, Telephony@Work, Signma Dynamics, Sunopsis, MetaSolv Software, Stellent, Hyperion, AppForge, SPL WorldGroup, Tangosol, LODESTAR Collectively > 200,000 database tables 500,000 million lines of code Now what?
5	Survey Non-Oracle DBMS Non-J2EE Application Server Apps user PeopleSoft JD Edwards Siebel eBusiness None/other Web technology J2EE .Net Other J2EE IDE JDeveloper Eclipse Other J2EE persistence ADF EJB EJB3 TopLink Hibernate
6	When I say “Fusion” I really mean… 1. Oracle Fusion - the next generation of application software Years away (2008? V1) Integration of the best features of all products 2. Fusion Middleware J2EE-based non-database technology to support Oracle Fusion (plus other stuff) OAS Parts of JDeveloper: ADF BC, ADF Faces Other stuff: BPEL, Business Rules tool, XML Publisher 3. Next versions of e-Business, PeopleSoft, JD Edwards using Fusion Middleware (not Retek)
7	Oracle Fusion Will be based on the e-Business data model Features of other packages will be migrated into e-Business. Migration path from PeopleSoft, JD Edwards, Siebel Impossible to automate Very expensive Ultimately essential Oracle cannot maintain all product stacks indefinitely. Fusion – V1 release scheduled for 2008 Will include the next major release of the e-Business suite using Fusion Middleware
8	Things that worry me about Fusion They really need to do a rewrite. They won’t do one. Would be cheaper with the right team SOA Focus It will just never, ever work. OO Folks have taken over architecture. Middle tier-centric code won’t work. Rewriting all PL/SQL to the middle tier was discussed. Rejected because of amount of code Not rejected because it was a bad idea
9	Getting there: Fusion Count on a significant conversion effort sometime within the next 5 years. New modules should be e-Business Move to Oracle DBMS Server-side PL/SQL Oracle Business Rules engine is in the DBMS. Move to Oracle Application Server Probably make life much easier
10	Fusion Middleware Definition Fusion OAS JDeveloper Developer Forms Reports Designer XML Publisher BPEL BAM Business Rules Engine Non-Fusion Application Express PL/SQL SQL Recently everything related to development is Fusion Middleware TopLink/Swing Integration EJB3
11	Two Goals of Fusion Middleware 1. Support Oracle Fusion Clear development path Tactical focus Strategic support HAS to work Limited scope 2. Support all J2EE development Market driven Lots of pieces Speculative Ill-defined scope
12	“Fusion Development Technology” (FDT) Not an Oracle term (but it should be) Subset of Fusion Middleware The technology used in Oracle Fusion This is what you really need to know.
13	Fusion Development Technology For the first time in Oracle’s history, development is THE critical success factor. At Collaborate ‘06, Charles Phillips’ keynote was “Fusion.” He never even mentioned the DBMS. Oracle is betting the farm on FDT. FDT is already good, and has all of the resources that it needs to become great. Will have a blank check for years to come
14	Fusion Development Technology Parts OAS J2EE application server First-rate product Mature Application Development Framework – Business Components (ADF BC) Persistence layer First-rate product Recently revamped ADF Faces Next generation UIX Somewhat proprietary Feels “new” Hard to go beyond framework BPEL Recent addition Hot standard for inter-system process Not sure where it fits Oracle Business Rules No idea what to do with this DBMS, PL/SQL, SQL
15	The Oracle DBMS It really is the best DBMS. But it is very complex. Licensing costs may be significant. Lower TCO Fusion will have lots of server-side code. Oracle Business Rules engine is in the DBMS. Server side code is MUCH better. Fusion code will migrate from middle tier to DBMS.
16	Case Study Batch routine Sales goaling
17	Getting There: The DBMS Can’t be avoided in the long run Huge shift from SQL Server You can use any operating system. Linux, Unix, Windows – all OK Start now It will take years. Wonderful engine SQL Lots of Oracle-specific features – learn them PL/SQL Server-side Oracle code – much more efficient Avoid server-side Java.
18	Oracle Application Server (OAS) J2EE application server Does not play well with MS Application Server No application server tech stacks interact well. Fusion will support other J2EE application servers. JDeveloper-to-OAS has single button deployment. Deploying to other J2EE application servers is annoying. Your life will be MUCH easier with OAS. Especially if doing custom deployment
19	Getting There: The Oracle Application Server Not such a big deal – can be avoided Unless you are using MS Application Server Better integration than other application servers Lowest TCO No finger pointing Lower deployment costs
20	ADF BC Persistence interface for Fusion Oracle alternatives TopLink Non-Oracle alternatives Hibernate – open source EJB EJB3 – supported in JDeveloper Very high-quality Proprietary framework Very little penetration outside of Oracle
21	Why use ADF BC? Lower probability of project failure Very rich product Very mature BC4J – V1 released in 2001 Rewritten several times Fusion will use it. Leading causes of J2EE project failure are hand-written persistence interfaces
22	Getting There: ADF BC Start now High learning curve Easy to misuse Oracle JDeveloper 10g Handbook (Roy-Faderman, Koletzke, Dorsey) Oracle JDeveloper 10g for Forms & PL/SQL Developers (Koletzke, Mills)
23	BPEL (Business Process Execution Language) Emerging standard Oracle implementation is very nice. Middle tier process flow language SOA inspired Makes great sense for inter-system flows Makes no sense for complex, local process flow
24	BPEL (“Things I don’t like about you”) 1. Not a repository-based technology The rules are the code No articulation/implementation independence 2. Middle tier-based If data intensive, lots of round trips 3. Not a good tool for describing a complex process 6,000 nodes for a single process description
25	BPEL (“Things I like about you”) Standards-based Good SOA integration Fairly easy to learn
26	Getting There: BPEL Wait and see. How will Oracle really use this? SOA: You may not even need it unless you are big. Process Flow: You have some time.
27	ADF Faces Rich (sort of) user interface Standards-based JavaServer Faces Proprietary extension of Faces Next generation UIX Not really mature Some quirks Evolving fast Just another tag library Hard to extend WYGIWYG (“What you get is what you get”) Still evolving
28	Getting There: Faces Start now Long learning curve Use with ADF Build a small project (or 3)
29	Oracle Business Rules Code-based ECA (Event-Condition-Action) architecture A solution in search of a problem I have no idea what to do with this. Has anyone tried to use this? Not a business rules product Articulation = implementation ECA engine builder Flawed meta-model Not enough reusability Wait to see what Oracle does with this in the applications
30	Getting There: Oracle Business Rules Wait and see. Not sure where this fits Focus resources elsewhere for now.
31	Playing well with others Other systems SOA, Web Services Can I keep my DBMS? Only for the short term Can I keep my J2EE Application Server? Yes, but I wouldn’t. Can I keep my “.NET” applications and non-J2EE application server? Yes, but it will be on its own box. J2EE and .Net can communicate using JNBridge. What about server-side MS code? Rewrite in PL/SQL
32	Non-Apps – Why should I care? Since I’m not an Oracle Applications customer, why should I care? This will be the best development platform on the planet. Great Oracle integration ADF BC is too good to ignore. Fusion will be a force in the industry and dominant within the Oracle development community.
33	Fusion Middleware - Conclusions A great (or will be soon) development environment Still evolving - all parts are not totally civilized. Seems weak for architects (but I am biased) Too good and big to ignore Will be the standard for all Oracle Applications (eBusiness, PeopleSoft, JD Edwards) Still evolving, so use “thick database” approach
34	Summary Oracle Fusion will be based on e-Business. Everyone else migrates. Must use: Oracle DBMS, J2EE stack Should use: OAS, ADF Avoid for now: BPEL, Oracle Business Rules
35	Share your Knowledge: Call for Articles/Presentations Submit articles, questions, … to IOUG – The SELECT Journal ODTUG – Technical Journal select@ioug.org pubs@odtug.com
36	Dulcian’s BRIM^® Environment Full business rules-based development environment For Demo Write “BRIM” on business card Includes: Working Use Case system “Application” and “Validation Rules” Engines
37	Contact Information Dr. Paul Dorsey – paul_dorsey@dulcian.com Dulcian website - www.dulcian.com